Definition and classification are critical features of scientific research and many areas of practice regardless of whether the phenomena in question are genuinely categorical or confinuous. An agreed upon definition of learning disability serves as an essenfial foundafion for research. Without an agreed upon definition that can be implemented reliably and validly, understanding the nature, causes, and best treatments for learning disabilities is unlikely. Although we have not yet achieved a reliable and valid definition and system of classification for learning disabitities, doing so now represents an attainable scientific objective because we now understand the fundamental problem that has prevented development of a reliable and valid definifion and classificafion system, and this understanding points the way to a potential solufion in the form of a model that incoporporates mulfiple sources of information in a theorefically grounded way. This project has three speciflc aims. The first is to further develop and test a constellation model of word-level developmental reading disability. This will be done through similafion, modeling of empirical relations among measures in the large-scale Florida PMRN database, and an empirical study of students with word-level reading disability. The second aim is to begin the process of extending the model to other potential learning disabilities in reading and writing. Doing so requires better understanding of the simultaneous development of oral language, reading, and writing skills, and this will be accomplished using a cohort-sequenfial longitudinal study. The third aim is to identify potenfial subtypes of reading and wrifing disabilities. When theorefically- motivated potential subtypes have been subject to rigorous empirical test, important knowledge has emerged such as support for a phonological form of developmental reading disability as a developmental difference but not for a surface form. Identifying potential subtypes of reading and writing disabilities will be accomplished by testing a priori proposed subtypes using a latent-variable regression-based approach applied to both reading and wrifing data collected in the previously menfioned empirical studies. RELEVANCE (See instructions): The proposed studies have the potential to provide a reliable and valid definition and system of classification for learning disabilities. Accomplishing this objective has the potenfial to accelerate our understanding ofthe nature of learning disabilities and to develop more effective means of prevention, eariy identification, and intervenfion. / PROJECT/PERFORIWANCE SITE{S) (if additional space is needed, use Project/